File size: 8,087 Bytes
a19f837
 
 
 
 
 
 
 
 
 
 
 
f23524f
a19f837
 
 
 
 
 
 
 
 
 
 
cd450eb
a19f837
 
 
 
 
 
 
c26cdc1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a19f837
 
 
 
 
 
 
 
 
 
c26cdc1
 
a19f837
 
c26cdc1
a19f837
 
 
 
c26cdc1
 
a19f837
 
c26cdc1
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
a19f837
 
 
 
c26cdc1
 
 
a19f837
 
c26cdc1
 
a19f837
c26cdc1
a19f837
 
c26cdc1
a19f837
 
 
 
 
 
 
c26cdc1
a19f837
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c446c5a
75a0ea0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c446c5a
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
import json      
import numpy as np      
import tensorflow as tf      
from tensorflow.keras import layers      
import gradio as gr      
import re    
import requests    
import math    
import sentencepiece as spm

# SentencePiece ๋กœ๋“œ (ํ† ํฌ๋‚˜์ด์ €๋ž‘ ํŠน์ˆ˜ ํ† ํฐ ID๋„ ๋™์ผํ•˜๊ฒŒ ์„ธํŒ…)
sp = spm.SentencePieceProcessor()
sp.load("ko_unigram4.model")

pad_id = sp.piece_to_id("<pad>")
if pad_id == -1: pad_id = 0
start_id = sp.piece_to_id("<start>")
if start_id == -1: start_id = 1
end_id = sp.piece_to_id("< end >")
if end_id == -1: end_id = 2
unk_id = sp.piece_to_id("<unk>")
if unk_id == -1: unk_id = 3

vocab_size = sp.get_piece_size()
max_len = 100

def text_to_ids(text):
    return sp.encode(text, out_type=int)

def ids_to_text(ids):
    return sp.decode(ids)

class RotaryPositionalEmbedding(layers.Layer):
    def __init__(self, dim):
        super().__init__()
        self.dim = dim
        inv_freq = 1.0 / (10000 ** (np.arange(0, dim, 2) / dim))
        self.inv_freq = tf.constant(inv_freq, dtype=tf.float32)

    def call(self, x):
        # x shape: (batch, heads, seq_len, depth)
        batch, heads, seq_len, depth = tf.unstack(tf.shape(x))

        t = tf.range(seq_len, dtype=tf.float32)  # (seq_len,)
        freqs = tf.einsum('i,j->ij', t, self.inv_freq)  # (seq_len, dim//2)

        emb_sin = tf.sin(freqs)  # (seq_len, dim//2)
        emb_cos = tf.cos(freqs)  # (seq_len, dim//2)

        # (seq_len, dim//2) -> (1, 1, seq_len, dim//2) to broadcast with x
        emb_cos = tf.reshape(emb_cos, [1, 1, seq_len, -1])
        emb_sin = tf.reshape(emb_sin, [1, 1, seq_len, -1])

        x1 = x[..., ::2]  # (batch, heads, seq_len, depth//2)
        x2 = x[..., 1::2]

        x_rotated = tf.stack([
            x1 * emb_cos - x2 * emb_sin,
            x1 * emb_sin + x2 * emb_cos
        ], axis=-1)  # shape (batch, heads, seq_len, depth//2, 2)

        x_rotated = tf.reshape(x_rotated, tf.shape(x))  # ๋‹ค์‹œ (batch, heads, seq_len, depth)
        return x_rotated

class GEGLU(tf.keras.layers.Layer):
    def __init__(self, d_model, d_ff):
        super().__init__()
        self.proj = layers.Dense(d_ff * 2)
        self.out = layers.Dense(d_model)
    def call(self, x):
        x_proj = self.proj(x)
        x_val, x_gate = tf.split(x_proj, 2, axis=-1)
        return self.out(x_val * tf.nn.gelu(x_gate))

class KeraLuxBlock(tf.keras.layers.Layer):
    def __init__(self, d_model, d_ff, num_heads=20, dropout_rate=0.1):
        super().__init__()
        self.ln1 = layers.LayerNormalization(epsilon=1e-5)
        self.mha = layers.MultiHeadAttention(num_heads=num_heads, key_dim=d_model // num_heads)
        self.dropout1 = layers.Dropout(dropout_rate)
        self.ln2 = layers.LayerNormalization(epsilon=1e-5)
        self.ffn = GEGLU(d_model, d_ff)
        self.dropout2 = layers.Dropout(dropout_rate)
        self.rope = RotaryPositionalEmbedding(d_model // num_heads)

    def call(self, x, training=False):
        x_norm = self.ln1(x)

        # MHA ์ฟผ๋ฆฌ, ํ‚ค์— RoPE ์ ์šฉ
        batch_size = tf.shape(x_norm)[0]
        seq_len = tf.shape(x_norm)[1]
        num_heads = self.mha.num_heads
        depth = (x_norm.shape[-1]) // num_heads

        # (batch, seq_len, d_model) -> (batch, num_heads, seq_len, depth)
        qkv = tf.reshape(x_norm, [batch_size, seq_len, num_heads, depth])
        qkv = tf.transpose(qkv, [0, 2, 1, 3])  # (batch, heads, seq_len, depth)

        # RoPE ์ ์šฉ (query, key ๋ชจ๋‘ ๋™์ผ x_norm ์‚ฌ์šฉํ•˜๋‹ˆ ๋‘˜ ๋‹ค ์ ์šฉ)
        q = self.rope(qkv)
        k = self.rope(qkv)

        # ๋‹ค์‹œ ์›๋ž˜ shape๋กœ
        q = tf.transpose(q, [0, 2, 1, 3])
        q = tf.reshape(q, [batch_size, seq_len, num_heads * depth])
        k = tf.transpose(k, [0, 2, 1, 3])
        k = tf.reshape(k, [batch_size, seq_len, num_heads * depth])

        # MHA ํ˜ธ์ถœ: query=k=v=x_norm, ํ•˜์ง€๋งŒ RoPE ์”Œ์šด q,k๋กœ ๋Œ€์ฒด
        attn_out = self.mha(query=q, value=x_norm, key=k, use_causal_mask=True, training=training)

        x = x + self.dropout1(attn_out, training=training)
        ffn_out = self.ffn(self.ln2(x))
        x = x + self.dropout2(ffn_out, training=training)
        return x
        
class KeraLux(tf.keras.Model):
    def __init__(self, vocab_size, seq_len, d_model, d_ff, n_layers, num_heads=20, dropout_rate=0.1):
        super().__init__()
        self.token_embedding = layers.Embedding(vocab_size, d_model)
        # pos_embedding ์ œ๊ฑฐ
        self.blocks = [KeraLuxBlock(d_model, d_ff, num_heads, dropout_rate) for _ in range(n_layers)]
        self.ln_f = layers.LayerNormalization(epsilon=1e-5)

    def call(self, x, training=False):
        seq_len = tf.shape(x)[1]
        x = self.token_embedding(x)
        for block in self.blocks:
            x = block(x, training=training)
        x = self.ln_f(x)
        logits = tf.matmul(x, self.token_embedding.embeddings, transpose_b=True)
        return logits

# ๋ชจ๋ธ ์ƒ์„ฑ & ๊ฐ€์ค‘์น˜ ๋ถˆ๋Ÿฌ์˜ค๊ธฐ
model = KeraLux(vocab_size=vocab_size, seq_len=max_len, d_model=160, d_ff=616, n_layers=6)
dummy_input = tf.zeros((1, max_len), dtype=tf.int32)  # ๋ฐฐ์น˜1, ์‹œํ€€์Šค๊ธธ์ด max_len
_ = model(dummy_input)  # ๋ชจ๋ธ์ด ๋นŒ๋“œ๋จ
model.load_weights("KeraLux3.weights.h5")
print("๋ชจ๋ธ ๊ฐ€์ค‘์น˜ ๋กœ๋“œ ์™„๋ฃŒ!")

def decode_sp_tokens(tokens):
    text = ''.join(tokens).replace('โ–', ' ').strip()
    return text

def generate_text_topkp_stream(model, prompt, max_len=100, max_gen=98, p=0.9, k=50, temperature=0.8, min_len=20):
    model_input = text_to_ids(f"<start> {prompt}")
    model_input = model_input[:max_len]
    generated = list(model_input)
    text_so_far = []

    for step in range(max_gen):
        pad_length = max(0, max_len - len(generated))
        input_padded = np.pad(generated, (0, pad_length), constant_values=pad_id)
        input_tensor = tf.convert_to_tensor([input_padded])
        logits = model(input_tensor, training=False)
        next_token_logits = logits[0, len(generated) - 1].numpy()

        if len(generated) >= min_len:
            next_token_logits[end_id] -= 5.0
        next_token_logits[pad_id] -= 10.0

        # ์˜จ๋„ ์ ์šฉ
        logits_temp = next_token_logits / temperature

        # 1. ํ™•๋ฅ  ๊ณ„์‚ฐ
        probs = tf.nn.softmax(logits_temp).numpy()

        # 2. Top-k ํ•„ํ„ฐ๋ง
        top_k_indices = np.argpartition(probs, -k)[-k:]
        top_k_probs = probs[top_k_indices]

        # 3. Top-p ํ•„ํ„ฐ๋ง (๋ˆ„์ ํ•ฉ ๊ณ„์‚ฐ์šฉ ์ •๋ ฌ)
        sorted_idx = np.argsort(top_k_probs)[::-1]
        top_k_indices = top_k_indices[sorted_idx]
        top_k_probs = top_k_probs[sorted_idx]
        cumulative_probs = np.cumsum(top_k_probs)

        # p ๋„˜๋Š” ๋ถ€๋ถ„ ์ž๋ฅด๊ธฐ
        cutoff = np.searchsorted(cumulative_probs, p, side='right') + 1

        filtered_indices = top_k_indices[:cutoff]
        filtered_probs = top_k_probs[:cutoff]

        # ํ™•๋ฅ  ์ •๊ทœํ™”
        filtered_probs /= filtered_probs.sum()

        # ์ƒ˜ํ”Œ๋ง
        next_token_id = np.random.choice(filtered_indices, p=filtered_probs)

        generated.append(int(next_token_id))
        next_word = sp.id_to_piece(int(next_token_id))
        text_so_far.append(next_word)

        decoded_text = decode_sp_tokens(text_so_far)

        if len(generated) >= min_len and next_token_id == end_id:
            break
        if len(generated) >= min_len and decoded_text.endswith(('.', '!', '?')):
            break

        yield decoded_text
    

history = ""

def chat(user_input):
    global history
    response = generate_text(user_input)  # ๋„ค ๋ชจ๋ธ ์ƒ์„ฑ ํ•จ์ˆ˜
    history += f"์‚ฌ์šฉ์ž: {user_input}\nKeraLux: {response}\n\n"
    return history

with gr.Blocks() as demo:
    gr.Markdown("### ๐Ÿ“Ÿ KeraLux Textbot\n๊ฐ„๋‹จํ•˜๊ณ  ๋น ๋ฅธ ๋Œ€ํ™”์šฉ ๋ด‡์ด์—์š”.\n")

    textbox = gr.Textbox(placeholder="๋ฉ”์‹œ์ง€๋ฅผ ์ž…๋ ฅํ•˜์„ธ์š”", lines=1)
    output_area = gr.Textbox(label="๋Œ€ํ™” ๊ธฐ๋ก", lines=20, interactive=False)
    
    textbox.submit(chat, inputs=textbox, outputs=output_area)
    textbox.submit(lambda: "", None, textbox)  # ์ž…๋ ฅ์ฐฝ ์ดˆ๊ธฐํ™”

demo.launch(share=True)